Re: Converting HbA1c to mg/dl

[Guest guest]

Another Ron Sebol post to the LC-DIABETES group that I thought you would

find interesting:

***********************************

Tom Rondo wrote:

>

> Could someone on the list reply with the math factor to convert the

> HgA1c to mg/dl ..... i.e. the scale used on our monitors.

It's 30 times A1c minus 60 if the lab has a normal range centered at

about 5.25.

Unfortunately, attempts to standardize the calibration of A1c test

methods (there are several different means of doing the measurement)

were started ten years ago and, for no rational reason, never made it

to the finish line. Several methods of standardization were proposed,

four as I recall, and none was selected as best and final. As a result,

the formula above may not apply to some labs even though it covers most

of them. Things get even more messy when a glucohemoglobin test is

confused with an A1c. In some cases, the only way to be truely sure how

to make the units conversion is to contact the lab and ask them the

question you have asked here on the list.

Sometimes a lab report will provide an A1c in percentage units and a

mg/dl calculated equivalent titled " average fasting blood glucose " , or

something similar. If you are lucky enough to have that dual

presentation, simply check to see if the mg/dl line is what you get with

the formula I cited. If so, that formula will continue to work but you

won't need it because that extra line eliminates the need to do a

conversion. I mention this because some people do not realize that the

two lines in their reports are really one in the same test. However,

since you ask how to make the conversion and probably knew what I just

stated, I doubt that you have the dual presentation form of report which

takes us back to contacting the lab.

If your question had nothing to do with your lab report and, instead,

you are trying to convert the DCCT or some other piece of DM

complication risk literature to mg/dl so you can better relate what they

found to what you see on your home meter, the formula above is

applicable.

Be warned that there are two different home testing meter scales in

vogue and both are in units of mg/dl; the industry savants really make

it hard to find something to admire. The plasma calibrated scale of the

Glucometer Elete is 1.12 that of the whole blood scaled One Touch, for

example. Further divergance derives from the home meter being off by

about 15% unless the linerar regression correction is applied and even

then it may be off by 5%. Add to that a 3% (1 sigma) random strip to

strip measurement noise and any one test is only good for about 8%

accuracy. Further, though well corellated, home meter readings are at

best only able to approximate an A1c. To make a best guess, take the

average of the three fasting readings before the three main meals of the

day. Multiply that by 1.07 and, if you have a whole blood scaled meter,

multiply that result by 1.12. That will be in the ball park of the A1c

but only accidentally exact. In the study that developed that

relationship the calculated result had a 93% corellation with the actual

A1c. It would not be at all unusual to be off 20 mg/dl from the A1c in

that three reading approximation method even if there were no meter

errors to contend with. The A1c is a continuous exponentially weighted

time average but the home readings are isolated spot measurements that

do not directly reflect an average of any kind. That is why you can only

approximate the lab test; home and lab are measuring things that are

intrinsically different. Related, but different.